Robert Macy wrote:



Interesting, the service/repair company claimed cap was one of the
major reasons for a unit to fail to run at all.

In electronic applications it is often possible to measure the cap in
situ, and determine its 'health' to a degree. My understanding is that
the standard failure mechanism is the esr starts increasing, the cap
dissipates more power, getting hotter the esr continues to increase,
and eventually POP. Starts as low as 10-30 milliohms then can climb up
passed 1-2 ohms depending. There are a whole bunch of DIY projects to
make inexpensive esr cap measurment instrumentation.
There may be certain brands or manufacturing date ranges where the
run caps are the most common point of failure. Most of my experience
is with older systems, in fact my house has two zones of 36 year old
Bryant A/Cs, and each year I expect to be the last for them, but they
keep soldiering on. I did blow a compressor run cap about 5 years ago.

But, my understanding of these oil-paper caps is that through thermal
cycling they eventually develop pinholes in the paper and the cap
goes from completely nominal to shorted within a single power line cycle.
So, unlike electrolytics that suffer slow loss of water in the electrolyte,
you DON'T get any warning or measurable degradation of the cap.

Jon